Selective funding for research is here to stay. The only question is whether to give top-rated departments an even bigger share, universities and colleges were told this week.

English ÖйúAƬ funding council chiefs have said they believe there should be no reduction in the present level of selectivity in handing out research money. But they have side-stepped the controversial question of whether a greater portion of the research cash pot should be devoted to areas that score well in the Research Assessment Exercise, leaving institutions to express their views first.

The ÖйúAƬ Funding Council for England's Funding Method for Research consultation paper gives universities and colleges until November 4. They also have until then to sift through the council's other research funding proposals, which it hopes to bring into operation from 1997/98.

They include the following: * Creating three new cost bands to set base research funding levels for subjects

ÖйúAƬ

* Adding a "policy weighting" to funding based on national needs and the international standing of particular research

* Maintaining the current system for counting research staff and students for the purposes of funding

ÖйúAƬ

* A new system for pump-priming to replace the "DevR" funding.

The council's views on selectivity have been informed by a study from consultants Segal Quince Wickstead (SQW) Ltd. The SQW study concluded that selectivity had allowed institutions to free gifted researchers for high-quality work of international importance. The policy also prompted institutions to give higher priority to planning and strategic development. The study found that institutions receiving only small sums from the council had obtained backing from other sponsors.

The consultation paper says the funding council plans to make final decisions on selectivity in December, following consultation and a thorough appraisal of the implications of the 1996 RAE results. But it adds that so far the council believes "that the degree of selectivity in distributing funds should be no less than now".

The council is also planning to set base values for the amount of funding provided for each subject. It proposes creating three broad cost bands in which subjects would be placed. The bands would be part-laboratory based; laboratory-based and clinical; and mainly library-based. Cost weights for each band would be calculated by the council using financial data from the sector.

ÖйúAƬ

A policy weighting might be added which would allow the funding council to influence the distribution of money between subjects. The consultation paper suggests that provision which is helping to meet national needs or is at the leading edge of international research in its field could be awarded extra funding through this system.

The paper says the council would draw on the expertise of peer and merit review bodies, including the research councils and learned societies, in developing judgements on national needs and international standing.

The council proposes to continue with its current method of counting numbers of research- active staff, research fellows, research assistants and research students, to gauge the volume of research activity in institutions. It has concluded that it would be too costly to administer a scheme where funding was shifted to reflect transfers of research staff between Research Assessment Exercises.

Money provided to establish collaborative research links between institutions and the commercial sector, the so-called "GR" element of funding, will remain unchanged. But the council is committed to reviewing "DevR" - pump-priming cash for institutions which have shown research potential but have not previously received significant research resources.

ÖйúAƬ

The paper suggests two options: - to reward institutions which have used DevR funds successfully and achieved higher RAE ratings this year than in the 1992 exercise; or to focus DevR on applied research only.